3.2 Determinants and their Properties
Every square matrix carries one number, its determinant. This section defines determinants of order two and three, evaluates them by expansion and by the properties that make the work short, and uses them to find the area of a triangle and to test whether three points are collinear.
Solve the pair of equations $a_1x+b_1y=c_1$, $a_2x+b_2y=c_2$ by elimination: multiply the first by $b_2$, the second by $b_1$ and subtract. You get $(a_1b_2-a_2b_1)\,x=c_1b_2-c_2b_1$. Everything depends on the single number $a_1b_2-a_2b_1$ built from the coefficients: if it is not zero there is exactly one solution, and if it is zero the two lines are parallel (no solution) or the same line (infinitely many). Mathematicians in seventeenth-century Japan and Germany noticed the same pattern for three equations in three unknowns, with a longer expression, and called it the determinant โ it determines whether the system has a unique solution.
The same number has a picture. The parallelogram with sides along $(a,b)$ and $(c,d)$ has area $|ad-bc|$, and that is why the area of a triangle can be written as a determinant of its vertices. In this section we define determinants of order $2$ and $3$, learn the properties that let you evaluate them in a few lines instead of grinding through six products, and apply them to areas and collinearity. In the next section the determinant becomes the key to the inverse of a matrix. JEE Main asks about this part of the unit in almost every paper โ a determinant to evaluate or factorise, a property like $|kA|=k^n|A|$ or $|AB|=|A||B|$, or an area.
Determinants of Order Two and Three
The determinant of a $1\times1$ matrix is its entry: $|[a]|=a$. The determinant of a $2\times2$ matrix is
$$|A|=\det A=\begin{vmatrix}a&b\\c&d\end{vmatrix}=ad-bc$$โ the product down the leading diagonal minus the product up the other diagonal. Only square matrices have determinants. We write $|A|$, $\det A$ or $\Delta$; the straight bars mean "determinant", not "absolute value" (a determinant can be negative).
For order three we reduce to order two. Delete a row and a column, and a $2\times2$ determinant is left over.
For a square matrix $A=[a_{ij}]$, the minor $M_{ij}$ of the entry $a_{ij}$ is the determinant left after deleting row $i$ and column $j$; its cofactor is $C_{ij}=(-1)^{i+j}M_{ij}$. The signs $(-1)^{i+j}$ form a chessboard pattern starting with $+$ in the top-left corner:
$$\begin{bmatrix}+&-&+\\-&+&-\\+&-&+\end{bmatrix}.$$For a $3\times3$ matrix, $|A|=a_{11}C_{11}+a_{12}C_{12}+a_{13}C_{13}$ (expansion along row $1$):
$$\begin{vmatrix}a_{11}&a_{12}&a_{13}\\a_{21}&a_{22}&a_{23}\\a_{31}&a_{32}&a_{33}\end{vmatrix}=a_{11}\begin{vmatrix}a_{22}&a_{23}\\a_{32}&a_{33}\end{vmatrix}-a_{12}\begin{vmatrix}a_{21}&a_{23}\\a_{31}&a_{33}\end{vmatrix}+a_{13}\begin{vmatrix}a_{21}&a_{22}\\a_{31}&a_{32}\end{vmatrix}.$$A $3\times3$ determinant can be expanded along any row or any column, and every expansion gives the same value:
$$|A|=\sum_{j=1}^{3}a_{ij}C_{ij}\ \ (\text{any row } i)\qquad=\qquad\sum_{i=1}^{3}a_{ij}C_{ij}\ \ (\text{any column } j).$$Written out in full, $|A|=a_{11}a_{22}a_{33}+a_{12}a_{23}a_{31}+a_{13}a_{21}a_{32}-a_{13}a_{22}a_{31}-a_{11}a_{23}a_{32}-a_{12}a_{21}a_{33}$.
Why. Expanding along row $1$ and multiplying out the three $2\times2$ determinants gives the six products in the box. Each product takes exactly one entry from every row and exactly one from every column, with the sign shown. Expanding along any other row or column produces the same six products with the same signs โ for example along column $1$: $a_{11}(a_{22}a_{33}-a_{23}a_{32})$ $-\,a_{21}(a_{12}a_{33}-a_{13}a_{32})$ $+\,a_{31}(a_{12}a_{23}-a_{13}a_{22})$, which multiplies out to the same six terms. So the choice of row or column is yours. $\blacksquare$ Use it: expand along the row or column with the most zeros โ a zero entry kills its whole term. The six-term formula can be remembered by Sarrus's rule: copy the first two columns to the right of the determinant, add the three products running down to the right and subtract the three running up to the right. (It works for order $3$ only.)
For $A=\begin{bmatrix}2&0&1\\3&-1&4\\5&2&6\end{bmatrix}$, the cofactor $C_{21}$ is
Properties of Determinants
Expanding a $3\times3$ determinant costs six products; with large or algebraic entries that is slow and error-prone. The properties below let you simplify first โ create zeros, take out common factors, spot equal rows โ and then expand. Each is stated for rows; by the first property, every one is equally true for columns.
Let $A$ be a square matrix of order $2$ or $3$.
- Transpose: $|A^{T}|=|A|$.
- Swap: interchanging two rows changes the sign of the determinant.
- Equal rows: if two rows are equal (or proportional), the determinant is $0$.
- Scaling a row: multiplying one row by $k$ multiplies the determinant by $k$. Hence a common factor of a row can be taken outside.
- Sum rule: if every entry of a row is a sum of two terms, the determinant splits into the sum of two determinants: $$\begin{vmatrix}a_1+b_1&a_2+b_2&a_3+b_3\\p_1&p_2&p_3\\q_1&q_2&q_3\end{vmatrix}=\begin{vmatrix}a_1&a_2&a_3\\p_1&p_2&p_3\\q_1&q_2&q_3\end{vmatrix}+\begin{vmatrix}b_1&b_2&b_3\\p_1&p_2&p_3\\q_1&q_2&q_3\end{vmatrix}.$$
- Row operation: adding a multiple of one row to another, $R_i\to R_i+kR_j$ ($i\ne j$), does not change the determinant.
- Triangular: the determinant of a triangular (in particular, a diagonal) matrix is the product of its diagonal entries.
Proofs. (1) Expanding $|A^{T}|$ along its first column is expanding $|A|$ along its first row (Theorem 3.7). (2) For order $2$: $\begin{vmatrix}c&d\\a&b\end{vmatrix}=cb-da=-(ad-bc)$. For order $3$, if rows $i$ and $j$ are swapped, expand along the third row: every minor of that row is a $2\times2$ determinant made from rows $i$ and $j$, now in the opposite order, so every minor โ and the total โ changes sign. (3) Swapping two equal rows leaves the matrix unchanged but changes the sign: $\Delta=-\Delta$, so $\Delta=0$. Proportional rows: take the factor out by (4) and two rows become equal. (4) and (5): expand along that row โ each term $a_{ij}C_{ij}$ is multiplied by $k$, or splits as $b_jC_{ij}+c_jC_{ij}$ (the cofactors of that row do not involve the row itself). (6) By (5), the new determinant is $|A|$ plus $k$ times a determinant whose rows $i$ and $j$ are equal, which is $0$ by (3). (7) Expand along the first column repeatedly: only the diagonal entry survives each time. $\blacksquare$
For square matrices $A$, $B$ of the same order $n$ ($n=2$ or $3$) and a number $k$:
$$|AB|=|A|\,|B|,\qquad |A^m|=|A|^m,\qquad |kA|=k^n|A|.$$In particular $|AA^{T}|=|A|^2$, $|-A|=(-1)^n|A|$, and every skew-symmetric matrix of odd order has determinant $0$.
Proof for order 2. With $A=\begin{bmatrix}a&b\\c&d\end{bmatrix}$, $B=\begin{bmatrix}p&q\\r&s\end{bmatrix}$, $AB=\begin{bmatrix}ap+br&aq+bs\\cp+dr&cq+ds\end{bmatrix}$ and
$$|AB|=(ap+br)(cq+ds)-(aq+bs)(cp+dr)=ad(ps-qr)-bc(ps-qr)=|A|\,|B|,$$since the terms $apcq$ and $brds$ cancel against $aqcp$ and $bsdr$. The same identity holds for order $3$ (a longer computation of the same kind). Then $|A^m|=|A|^m$ by repeating it. $kA$ multiplies each of the $n$ rows by $k$, so by property (4) the determinant is multiplied by $k$ once per row: $k^n$. Finally, if $A^{T}=-A$ with $n$ odd, $|A|=|A^{T}|=|-A|=(-1)^n|A|=-|A|$, so $|A|=0$. $\blacksquare$ Warning: there is no such rule for sums โ in general $|A+B|\ne|A|+|B|$.
If the entries of one row are multiplied by the cofactors of a different row and added, the result is $0$: for $i\ne k$,
$$a_{i1}C_{k1}+a_{i2}C_{k2}+a_{i3}C_{k3}=0,$$and similarly for columns. (With $i=k$ the same sum is $|A|$.)
Proof. The sum is exactly the expansion, along row $k$, of the determinant obtained from $A$ by replacing row $k$ with a copy of row $i$ โ the cofactors of row $k$ do not see what is in row $k$. That determinant has two equal rows, so it is $0$ by Theorem 3.8(3). $\blacksquare$ This small fact is what makes the adjoint and the inverse work in Section 3.3.
A square matrix $A$ is singular if $|A|=0$ and non-singular if $|A|\ne0$. By Theorem 3.9, a product $AB$ of square matrices is non-singular exactly when both $A$ and $B$ are.
- Look first. Two equal or proportional rows (or columns)? A row of zeros? Then $\Delta=0$ at once. Triangular? Multiply the diagonal.
- Take out common factors from rows and columns (property 4).
- Make zeros with operations $R_i\to R_i+kR_j$ or $C_i\to C_i+kC_j$ (property 6) โ adding all rows to the first, $R_1\to R_1+R_2+R_3$, is useful when every column has the same sum.
- Expand along the row or column with the most zeros, minding the chessboard signs.
- Factor theorem for algebraic entries: if putting $x=a$ makes two rows equal, then $(x-a)$ is a factor of $\Delta(x)$; compare degrees and leading coefficients to find the rest.
Take the rows $R_1=(3,1)$ and $R_2=(1,2)$, so $\begin{vmatrix}3&1\\1&2\end{vmatrix}=5$. Apply $R_1\to R_1+2R_2$, which gives $R_1=(5,5)$. What is the new determinant?
$A$ is a $3\times3$ matrix with $|A|=5$. Then $|-2A|$ equals
Area of a Triangle and Collinearity
Widget 1 suggests that $|ad-bc|$ is the area of the parallelogram spanned by $(a,b)$ and $(c,d)$. Half of that parallelogram is a triangle, and this gives a determinant formula for the area of any triangle from the coordinates of its corners.
The area of the triangle with vertices $P_1(x_1,y_1)$, $P_2(x_2,y_2)$, $P_3(x_3,y_3)$ is $\tfrac12|\Delta|$, where
$$\Delta=\begin{vmatrix}x_1&y_1&1\\x_2&y_2&1\\x_3&y_3&1\end{vmatrix}=x_1(y_2-y_3)+x_2(y_3-y_1)+x_3(y_1-y_2).$$$\Delta\gt0$ when $P_1\to P_2\to P_3$ runs anticlockwise and $\Delta\lt0$ when it runs clockwise. The three points are collinear if and only if $\Delta=0$, and the line through $(x_1,y_1)$ and $(x_2,y_2)$ is
$$\begin{vmatrix}x&y&1\\x_1&y_1&1\\x_2&y_2&1\end{vmatrix}=0.$$Proof. Apply $R_2\to R_2-R_1$ and $R_3\to R_3-R_1$ (property 6) and expand along the third column:
$$\Delta=\begin{vmatrix}x_1&y_1&1\\x_2-x_1&y_2-y_1&0\\x_3-x_1&y_3-y_1&0\end{vmatrix}=\begin{vmatrix}p&q\\r&s\end{vmatrix}=ps-qr,$$where $(p,q)=P_2-P_1$ and $(r,s)=P_3-P_1$ are the two sides from $P_1$. Write them in polar form, $(p,q)=u(\cos\alpha,\sin\alpha)$ and $(r,s)=v(\cos\beta,\sin\beta)$, with $u$, $v$ their lengths. Then $ps-qr=uv(\cos\alpha\sin\beta-\sin\alpha\cos\beta)$ $=uv\sin(\beta-\alpha)$. The angle of the triangle at $P_1$ is $|\beta-\alpha|$ (or $360^\circ$ minus it), so the familiar formula "area $=\tfrac12\times$ two sides $\times$ sine of the included angle" gives area $=\tfrac12 uv|\sin(\beta-\alpha)|=\tfrac12|\Delta|$. The sign of $\sin(\beta-\alpha)$ is positive exactly when $P_3$ is reached from $P_2$ by turning anticlockwise about $P_1$. Three points are collinear exactly when the triangle has zero area. Finally, a point $(x,y)$ lies on the line through $(x_1,y_1)$ and $(x_2,y_2)$ exactly when the three points are collinear. $\blacksquare$
Fix $A(0,0)$ and $B(4,0)$, and let $C$ slide along the line $y=3$ from $x=-3$ to $x=6$. What happens to $\Delta$ for the triangle $ABC$?
1. $|kA|=k^n|A|$, not $k|A|$. Here $n$ is the order. For order $3$: $|2A|=8|A|$ and $|-A|=-|A|$. Multiplying one row by $k$ multiplies by $k$ only once.
2. Minor or cofactor? $C_{ij}=(-1)^{i+j}M_{ij}$. The signs follow the chessboard $+\,-\,+$; forgetting them is the commonest error in expansions.
3. No sum rule for matrices. $|A+B|\ne|A|+|B|$ in general. The sum rule (property 5) splits one row at a time.
4. Area questions have two cases. "Area $=4$" means $\Delta=8$ or $\Delta=-8$. And an area is never negative โ take $\tfrac12|\Delta|$.
5. Which row operations are free? Only $R_i\to R_i+kR_j$. The operation $R_i\to kR_i+R_j$ multiplies the determinant by $k$; swapping rows changes its sign. Sarrus's diagonal rule works for $3\times3$ only.
The points $(1,2)$, $(3,k)$ and $(5,8)$ are collinear when $k$ equals
Exercises
Use the properties of Theorem 3.8 wherever they shorten the work; state which property you use.
- 1 Evaluate $\begin{vmatrix}7&-3\\4&2\end{vmatrix}$.
- 2 Evaluate $\begin{vmatrix}\cos\theta&-\sin\theta\\\sin\theta&\cos\theta\end{vmatrix}$.
- 3 Evaluate $\begin{vmatrix}1&2&-1\\3&0&2\\-2&1&4\end{vmatrix}$.
- 4 Evaluate $\begin{vmatrix}0&5&-2\\-5&0&3\\2&-3&0\end{vmatrix}$ without expanding.
- 5 Find $x$ if $\begin{vmatrix}2x&5\\8&x\end{vmatrix}=\begin{vmatrix}6&-2\\7&3\end{vmatrix}$.
- 6 For $A=\begin{bmatrix}1&-2&3\\4&0&-1\\2&5&6\end{bmatrix}$, find the minors and cofactors of the entries of row $2$, and hence $|A|$.
- 7 Find the real $x$ with $\begin{vmatrix}x&0&1\\2&x&0\\0&1&x\end{vmatrix}=10$.
- 8 Evaluate $\begin{vmatrix}2&7&-1\\0&-3&5\\0&0&4\end{vmatrix}$.
- 1 Without expanding, show that $\begin{vmatrix}2&3&5\\4&6&10\\1&7&8\end{vmatrix}=0$ and $\begin{vmatrix}11&12&13\\12&13&14\\13&14&15\end{vmatrix}=0$.
- 2 Show that $\begin{vmatrix}1&a&b+c\\1&b&c+a\\1&c&a+b\end{vmatrix}=0$.
- 3 Using Example 2(b), evaluate $\begin{vmatrix}1&1&1\\1&2&4\\1&4&16\end{vmatrix}$ without expanding.
- 4 Show that $\begin{vmatrix}a+1&a&a\\a&a+1&a\\a&a&a+1\end{vmatrix}=3a+1$, and find its value for $a=4$.
- 5 $A$ is $3\times3$ with $|A|=3$. Find $|2A|$, $|A^{T}|$, $|A^3|$ and $|-A^{T}A|$.
- 6 Show that every $3\times3$ skew-symmetric matrix has determinant $0$. Is this true for $2\times2$ skew-symmetric matrices?
- 7 $A$ and $B$ are $3\times3$ with $|A|=2$ and $|B|=-3$. Find $|A^2B^{T}|$ and $|3AB|$.
- 8 Prove the sum rule (property 5) for a $3\times3$ determinant by expanding along the split row.
- 9 Show that $\begin{vmatrix}x&2&2\\2&x&2\\2&2&x\end{vmatrix}=(x-2)^2(x+4)$ and solve $\Delta=0$.
- 1 Find the area of the triangle with vertices $(1,0)$, $(6,0)$, $(4,3)$.
- 2 Find the area of the triangle with vertices $(-3,-2)$, $(2,1)$, $(-1,4)$.
- 3 The triangle with vertices $(k,3)$, $(2,-1)$, $(5,2)$ has area $6$ square units. Find all possible $k$.
- 4 Show that $(a,\,b+c)$, $(b,\,c+a)$, $(c,\,a+b)$ are collinear for all $a$, $b$, $c$.
- 5 Use a determinant to find the equation of the line through $(2,-1)$ and $(-1,5)$.
- 6 Find $k$ if $(k,-1)$, $(2,1)$, $(4,5)$ are collinear.
- 7 Find the area of the quadrilateral with vertices $(1,1)$, $(5,2)$, $(4,6)$, $(-1,4)$, taken in that order.
- 8 A surveyor marks the corners of a triangular plot at $(0,0)$, $(120,30)$ and $(40,90)$, measured in metres. Find the area of the plot.
- 1 Solve $\begin{vmatrix}x&3&4\\3&x&4\\3&4&x\end{vmatrix}=0$ and find the product of the roots.
- 2 Show that $\begin{vmatrix}a&b&c\\b&c&a\\c&a&b\end{vmatrix}=3abc-a^3-b^3-c^3$, and evaluate it for $a=1$, $b=2$, $c=3$.
- 3 $A$ is a $3\times3$ matrix with $AA^{T}=4I$ and $|A|\gt0$. Find $|A|$.
- 4 For $A=\begin{bmatrix}1&2&3\\0&1&4\\0&0&2\end{bmatrix}$, find $|2A^2|$.
- 5 Show that $\begin{vmatrix}x&x+1&x+2\\x+3&x+4&x+5\\x+6&x+7&x+8\end{vmatrix}=0$ for every $x$.
- 6 Find the set of points $P(x,y)$ for which the triangle with vertices $P$, $(1,1)$, $(4,5)$ has area $5$ square units.
E1. JEE Main ยท MCQ $A$ is a $3\times3$ matrix with $|A|=2$. Then $\big|3A^{T}A^2\big|$ equals (A) $54$ (B) $216$ (C) $72$ (D) $24$
E2. JEE Main ยท Numerical The points $(2,3)$, $(5,k)$ and $(k,7)$ are collinear. Find the sum of the squares of all possible values of $k$. Give the answer to the nearest integer.
E3. JEE Main ยท MCQ If $\begin{vmatrix}a&b&c\\p&q&r\\x&y&z\end{vmatrix}=5$, then $\begin{vmatrix}2a&2b&2c\\p+3a&q+3b&r+3c\\-x&-y&-z\end{vmatrix}$ equals (A) $10$ (B) $-30$ (C) $-10$ (D) $30$
E4. JEE Main ยท Numerical Find the number of singular $2\times2$ matrices whose entries all belong to $\{0,1,2\}$. Give the answer to the nearest integer.
E5. JEE Main ยท MCQ The triangle with vertices $(-1,0)$, $(3,0)$ and $(a,b)$ has area $6$ square units, and $(a,b)$ lies on the line $y=x+2$. The possible values of $a$ are (A) $1$ and $-5$ (B) $1$ and $5$ (C) $-1$ and $5$ (D) $3$ and $-3$
E6. JEE Main ยท Numerical Find the sum of the squares of the real roots of $\begin{vmatrix}x&1&2\\1&x&2\\1&2&x\end{vmatrix}=0$. Give the answer to the nearest integer.
Research question. Farmers, forest departments and town planners now measure land from GPS coordinates instead of tapes and chains. How accurately does the determinant (triangle) formula give the area of an irregular field, and how much do small GPS errors matter?
Why it matters. Land records, crop-insurance claims and compensation for land taken for roads are all paid per hectare. A field mapped as a polygon is split into triangles from one corner, and each triangle's area is half a $3\times3$ determinant of its corners (Theorem 3.11). Consumer GPS receivers are typically accurate to a few metres, so it is worth knowing how errors in the corners turn into errors in the area.
What to do.
- On an online map, choose a local park, playground or field with a clear boundary. Read off the coordinates (in metres, from a corner you choose as origin) of $6$ to $10$ corner points, going round anticlockwise.
- Split the polygon into triangles from the first corner and compute each area with a determinant. Add them, and compare with the shoelace form $\tfrac12\big|\sum(x_iy_{i+1}-x_{i+1}y_i)\big|$.
- Simulate GPS error: add a random shift of up to $\pm3$ m to each coordinate and recompute the area. Repeat $20$ times and record the spread of the results.
- Repeat step 3 for a field twice as long and twice as wide. Does the percentage error grow or shrink?
Tools: An online map with coordinates, a spreadsheet (RAND() for the random shifts), or a short Python script.
Hand in: Your list of corners, the computed area, the table of simulated areas with their range, and a paragraph advising a farmer how many corner points to record and how accurate the result will be.
Going further. Using $\Delta=x_1(y_2-y_3)+x_2(y_3-y_1)+x_3(y_1-y_2)$, show that moving one corner by $(\delta x,\delta y)$ changes $\Delta$ by an amount that does not depend on that corner's own position. What does this say about which corners deserve the most careful measurement?
๐งช Self-check quiz
Q1. $\begin{vmatrix}4&-2\\3&1\end{vmatrix}$ equals
Q2. If two rows of a determinant are interchanged, its value
Q3. $A$ is a $2\times2$ matrix with $|A|=3$. Then $|4A|$ is
Q4. The points $(0,0)$, $(2,4)$ and $(3,k)$ are collinear when $k$ is
Q5. For every $3\times3$ skew-symmetric matrix $A$, $|A|$ is